Day 2 of Linear Systems (Replaces 7.
2)
• What is a matrix? Determinant? Trace?
• Be able to convert a matrix-vector equation into a system of equations and vice versa.
• Be able to write the solution to the homogeneous system of equations in vector form
(also- Equilibrium solutions)
• Recall how to write lines in 2d, 3d using vector format.
• Understand the vector ekt v, especially geometrically.
• If ansatz if x(t) = eλt v, what must be true for x0 = Ax?
Class Notes
First some useful notation.
Key Definition: A system of equations can be written in matrix-vector form as shown
below (this is a definition)
ax + by = e a b x e
⇔ =
cx + dy = f c d y f
Similarly, we could extend this to three variables (this is just to show you what it would
look like):
a11 x1 + a12 x2 + a13 x3 = b1 a11 a12 a13 x1 b1
a21 x1 + a22 x2 + a23 x3 = b2 ⇔ a21 a22 a23 x2 = b2
a31 x1 + a32 x2 + a33 x3 = b3 a31 a32 a33 x3 b3
More Definitions...
A matrix is simply an array of numbers, and the size of a matrix is defined as the number
of rows × the number of columns (similar to a spreadsheet, rows always come first). We
identify elements of the array by locating the row and column.
In the examples above, we have a 2 × 2 matrix and a 3 × 3 matrix. A vector can also be
described as a 2 × 1 matrix (or 3 × 1).
The definitions above gives meaning to matrix-vector multiplication. For example,
1 2 1 1(1) + 2(2) 5
= =
−1 3 2 −1(1) + 3(2) 5
1
In the previous lesson, we reviewed the determinant. The trace of a matrix is the sum
of the diagonal elements of a matrix (from upper left element to the lower right element).
For example,
1 2
A= det(A) = 4 − 6 = −2 Tr(A) = 1 + 4 = 5
3 4
Review: Solving a System of Equations
In solving a linear system of equations, there are three (and only three) possible outcomes:
(i) Exactly one solution (intersecting lines), (ii) No Solution (parallel lines), (iii) an infinite
number of solutions (the same line).
We can solve the system using Cramer’s Rule:
e b a e
ax + by = e f d c f
⇒ x= y=
cx + dy = f a b a b
c d c d
This works as long as the determinant in the denominator is not zero. If that is the case,
then we either have parallel lines or the same line.
Example
Solve the system:
1 2 x 0
=
2 4 y 0
SOLUTION: The determinant is 0, and we see these are the same line. Therefore, any (x, y)
satisfying x + 2y = 0 will solve the system- We want to represent the line in vector form.
x −b
Shortcut: The line ax + by = 0 can be represented by the vector =t
y a
x −2
To solve the previous system, we should provide the line: =t
y 1
Lines in Vector Form
Recall from Calculus III: A line in two or three dimensions can be defined if it goes through
point p and goes in the direction q as (in parametric vector form):
p~ + t~q −∞<t<∞
2
So, for example, the line going through the point (1, 2, 3) in the direction of h1, −1, 1i can
be written as:
1 1 1+t
2 + t −1 = 2 − t
3 1 3+t
Class example: What does this look like:
1 1
2 + et −1 −∞<t<∞
3 1
SOLUTION: It is a ray rather than a line. As t → −∞, et → 0 (the line goes to (1,2,3)),
then as t increases, we move farther and farther in the direction given.
Systems of DEs and Matrices
Definition: An autonomous system of first order linear differential equations is a system
of the following form- All 3 forms of the notation are used at some point:
x01 = ax1 + bx2
0
x1 a b x1
⇔ = ⇔ x0 = Ax
x02 = cx1 + dx2 x02 c d x2
(Note: Autonomous means that t is not explicitly part of the DEs)
Definition: A solution to the system is a parametric function that satisfies the given
relationship.
To find the equlibrium solution, set the derivatives equal to zero and solve the corre-
sponding system- you might note that the zero vector is always an equilibrium solution, but
there may be more.
Example
−t 1
Show that x(t) = e solves the system:
2
x01
3 −2 x1
=
x02 2 −2 x2
For the derivatives, remember that x(t) is in parametric form, where x1 (t) = e−t and x2 (t) =
2e−t . We differentiate component-wise, so
0
−e−t
0 x1 (t) −t 1
x (t) = 0 = −t = −e
x2 (t) −2e 2
(Note that we could have simply differentiated the exponent in front, since that was the only
term that involved t).
3
Now we want to see if this is the same as Ax. Since e−t is a scalar, I’m able to factor it
out:
3 −2 x1 −t 3 −2 1 −t 3−4 −t −1 −t 1
Ax = =e =e =e = −e
2 −2 x2 2 −2 2 2−4 −2 2
We see that indeed, x0 (t) is the same as Ax(t).
Towards a General Solution
Suppose we have the system of differential equations:
x01 = ax1 + bx2
x02 = cx1 + dx2
with the ansatz below, where we stress that v 6= 0:
x1 (t) λt v1
x(t) = =e
x2 (t) v2
(Note that λ, v1 , v2 are all unknown in the ansatz as of now). What must be true of λ, v1 , v2 ?
Let’s substitute our solution into the system and see what we get. First, the derivative
is simple: 0
0 x1 (t) λt v1
x (t) = = λe
x02 (t) v2
And putting this into the system and simplifying:
x01 = ax1 + bx2 λeλt v1 = aeλt v1 + beλt v2 av1 + bv2 = λv1
⇒ ⇒
x02 = cx1 + dx2 λeλt v2 = ceλt v1 + deλt v2 cv1 + dv2 = λv2
From this, we get the really important set of equations that we want to remember. In order
for λ and v to be in the solution x(t), we need:
(a − λ)v1 +bv2 = 0
cv1 +(d − λ)v2 = 0
Now, if we were solving this using Cramer’s Rule, the determinant
a−λ b
c d−λ
would be in the denominator, and if this were non-zero, the ONLY SOLUTION is the zero
vector. We don’t want that- We want v to be NOT zero, so we must find λ so that the
determinant IS zero. That is our first condition: Find λ so that
a−λ b
=0
c d−λ
4
Once we find λ, we go back to the system we were looking at, and substitute our new value
of λ into the system:
(a − λ)v1 +bv2 = 0
cv1 +(d − λ)v2 = 0
Which is guaranteed to represent two multiples of the same line (that’s how we defined λ).
We’ll continue this discussion in our next lecture.
Homework (to replace 7.2)
2t 1
1. What will the graph of e be (where t is any real number).
2
2. Verify that x1 (t) below satisfies the DE below.
0 1 1 3t 1
x = x, x1 (t) = e
4 1 2
3. Each system below is nonlinear. Solve each by first writing the system as dy/dx.
x0 = y(1 + x3 ) x0 = 4 + y 3 x0 = 2x2 y + 2x
(a) (b) (c)
y 0 = x2 y 0 = 4x − x3 y 0 = −(2xy 2 + 2y)
(Note: Some of these may be exact.)
a−λ b
4. For each matrix below, compute the determinant . Your determinant
c d−λ
should be an expression in λ.
1 1 3 −2 1 −2
(a) (b) (c)
4 1 2 −2 3 −4
a b a−λ b
5. Show that if the matrix A = , then the determinant can be
c d c d−λ
expressed as:
λ2 − tr(A)λ + det(A)
where tr(A) is the trace of A and det(A) is the determinant of A.
3 −2
6. Suppose λ = 2. Using the matrix , solve the following system for v1 , v2 . Be
2 −2
sure to write your answer in vector form.
(a − λ)v1 +bv2 = 0
cv1 +(d − λ)v2 = 0
5
3 −2
7. Suppose λ = −1. Using the matrix , solve the following system for v1 , v2 .
2 −2
Be sure to write your answer in vector form.
(a − λ)v1 +bv2 = 0
cv1 +(d − λ)v2 = 0